 From London, England, it's the Cube. Covering Discover 2016 London. Brought to you by Hewlett Packard Enterprise. Now, here's your host, Dave Vellante. Get Excel London at the docks, where it's 27 degrees outside. I feel like I'm back in Boston, but this is HPE Discovering. This is the Cube, the worldwide leader in live tech coverage. Dom Wild is here. He's the vice president and general manager of the data center networking group at Hewlett Packard Enterprise. Mark Smith is the senior vice president of sales and business development at Arista. And Richard Haig is the head of delivery enablement at Paddy Power Betfair. Gentlemen, welcome to the Cube. Good to see you again, Dom. Thank you. So we've been hearing a lot about the partnership with Arista. You guys have been making a lot of noise there. Andy Bechtelstein was on the stage yesterday in a suit. Last time I saw him, he was in sort of street clothes. And it's high. Dresses up for the European audience. I love it. But Dom, set up the sort of where you're at with the networking group, generally and specifically what's going on with Arista. Yeah, so with networking, we've seen a real segmentation in the market, really around a set of customers who are looking at sort of more traditional infrastructures and need sort of breadth and depth of feature sets and more sort of traditional legacy kind of environments. But the real shift that's happening is obviously towards software defined infrastructure for the cloud. And Arista is obviously a leader in that space with incredible growth and really impressive growth that Mark's been driving. And our philosophy is really to go and deliver best of breed solutions and open solutions for our customers. And so it's really a sort of natural partnership for us to partner with Arista, to deliver on the promise of software defined and cloud. Help customers transform. And Mark, the nature of the partnership is what? HPE is reselling your system and you've got solutions together? Yeah, they are. I mean, it's a big partnership for us, right? We've been a huge cloud networking company. We've won all the major cloud customers, R1. And now we want to take that same technology which customers want through HPE to the enterprise. Right, I mean, we're going to come back and talk about how you've competed with Cisco. You've done a great job of that and you've crushed it in the cloud service provider space. But before we go there, Richard, set up Paddy Power, Betfair, what you guys do, what the company's all about. Yep, sure. So Paddy Power, Betfair formed from a merger in February this year from Paddy Power and from Betfair, both large online gaming companies and they've come together to form one of the world's largest online gaming companies. So it's a very big, technically driven business. And this is why, of course, we're interested in some of the bits and pieces. So the nature of that business is a lot of data, a lot of analytics behind that and very competitive. Of course, we're in the UK, which is betting central. So what's the market like for you guys right now? So the market predominantly was in UK Ireland, but it's growing more and more global. We're pushing out into further regulated markets around the globe and that brings its own challenge in terms of scaling the infrastructure and keeping pace with the development. It's all very digitally driven, there is a retail segment, but even that we're looking to push more and more through the technology side of it. And that just gives us more and more of a challenge in building up and supporting that infrastructure to keep those features and that functionality ever more present. And you're a customer of the HPE, ERISTA sort of partnership solution, talk about, paint a picture of what the infrastructure is like, but specifically the networking piece. Sure, so we went probably about a year, a year and a half ago, we went out to market looking for the various providers that would make up the stack that we would consume. It's a large open stack installation, so we had a lot of freedom to choose the various components to fit together. And through that stack, we ended up choosing independently of each other at the time, HPE and ERISTA to provide that networking layer and to provide that x86 compute layer. The real driver for us was finding supplies to offer a rich API and allowed us to programmatically interface with them. So our dream for this new infrastructure was to define everything as code. So we wanted our development community to be able to make changes to the infrastructure exactly as they would to any public cloud provider they might use, but in our own private cloud space. So the real importance of that programmatic API driven infrastructure was absolutely paramount. So to sense, Mark, this is kind of a similar, I mean, you guys, like I said before, crushing cloud service providers. This is sort of an example where you've got a company that looks like a CSP, but is doing their own thing. Describe that a little bit. Well, I think, I think Betfair is a lot like many companies is they actually are looking at building internal clouds and eventually they'll probably go to a hybrid cloud model and they want to resemble what the clouds are doing, which is basically automate provision in an automated way and run really reliable networks that they can run at scale. So talk a little bit more Tom about the portfolio and where our risk to fits in. Obviously it's high end. And then I want to understand how you guys have been so successful competing against Cisco specifically, but go ahead. So the partnership officially kind of got off the ground and launched November 1st. We actually announced back in September, but we actually started shipping product November 1st. One of the questions that people have asked me is, well, you had a portfolio, so now you're replacing something. The answer is back to the segmentation. We've had our flex fabric products for many, many years now and those continue forward. We'll continue to invest there, really aligned that sort of that more traditional use case. Arista fits in beautifully in the growth segment for software defined in cloud. And we've announced and made a very strong statement that Arista is our preferred partner in that space. And then there's the other segment, which is sort of the mega scale sort of service providers who are going after this kind of white box model and we have a portfolio called Alto line of white boxes with some choice of OS there. So we're looking at that segmentation. I mean, it's not hard and fast. There's gray areas between there, but really the growth area for the future is around cloud, it's around convergence. So that's where the partnership comes in. We're really excited about it. Our customers are incredibly excited about it. I think Richard and his team are of validation of this thinking around how we're doing this. I, Richard, you're also using the Nuage products as well for network virtualization, OpenStack. These are all things that HPE offers. So Arista, Nuage, OpenStack, these are all the leading solutions and the best-of-breed solutions that we can bring to bear alongside our sort of server and storage operations. Mark, J.Sri Ullal was the first to ever explain to me what was happening with network traffic, collapsing, going from North-South hierarchical to East-West. Can you describe sort of the mega trends that have been driving momentum in the marketplace? Well, I think that a few things that have been talked about already is open. So most of our customers and the cloud providers want an open network. They're tired of being locked in to fabrics that where the vendor controls their destiny and they also raise their prices over time and they also, at their whim, get to end-of-life products, rip them out and replace and most companies today, the CIO is looking to cut operational costs and they can't do that with the traditional lock-in of black box mentality. So I think open is a huge trend in what's happening in data center. The second thing is people are looking for a lot more automation. Software-defined networking is all about automating standard processes where humans are involved and can cause errors and problems and 60% of your IT staff or networking team is spent working on problems. By having it automated, I can eliminate human errors which are probably, I don't know, 90% of the problems that you have in a network are created because somebody creates a wrong network they type in something wrong, they do a DNS, any type of query that they do incorrectly creates outages and so when I can automate a network, I can eliminate that human factor and I can run a much more reliable network at scale, the way that Betfair is doing. So let's talk about that, Richard. How much of what Mark said sort of resonates with you and fed into the decision for you guys to go with HB University? He's spot on, so one of the big drivers is can we go faster? Our CEO is famous for saying, great, good solution, if I gave you more money could you do it more quickly and being able to automate that network layer and be able to immutably deploy that network layer certainly helps us do that. So it's spot on as well about the number of errors. So when you look at things like large East West firewall rule sets as they build up over five, 10 years of the history of the business, you end up with these huge monolithic rule sets that are very difficult to understand. You might want to make a simple change for a new product that's coming online and it has a knock on consequence that gives you an outage and you really have to spend a lot of time discovering why. Because we've been able to take the network and make it software defined and we've been able to put it into an immutable model so rather than building on top of the rule set that is there all the time we blow away a subnet, we blow away the AC rules for that app and we build them from scratch every time. We end up with a much more segmented way of putting all these rule sets in place and all of these apps interacting and each independent application owner has a very good understanding of the rules and the networking around their specific app rather than having to pin back to a central networking team who finds it impossible to keep it all in their head about all of this. And you're saying they can start with a clean slate pretty much every time they want to roll out a new project or a new app or. Absolutely and because it's all defined as code as well our development community, this is the language, this is how they speak so rather than having to interact with various GUIs or various products or portals we now allow them to check in a YAML code defining their intent for the network or for the firewall ruling surrounding their application that's checked in and it goes through the same source controlled testing pipeline behavior that all of the rest of our code does the outcome just happens to be infrastructure so it really has sped up our ability to provide that infrastructure but also given massive flexibility to the development teams about how they consume it. You're describing what people refer to as the dev ops environment more so than ops dev, in other words sort of force fitting your operations people into a development role, is that true? Yeah absolutely, my guiding kind of light was I want to make this so easy that they don't want to use AWS because this is easier and so we're really trying to provide it dev centric so that it's everything that the devs want to do we provide them a way to do that under their own control. The flip side of this as well is we get fantastic feedback from our security audit guys because now everything is as code and source controlled and historically evident so for the guys doing audits they have absolute granularity over everything that has been changed and the version of those changes going all the way back to the start so we've got a very happy development community who is able to have a fast interactive approach to their infrastructure a very happy security entity who has absolute visibility over everything that's happened so it's a real win-win across this. Can you paint a even deeper broader picture of your networking infrastructure? What's it look like under the covers? Sure, so the biggest change we've made is we've gone from a core networking topology to spine leaf and that spine leaf has allowed us to have much lower latency across the network design so we're two hops minimum away from just about anything but then in the context we've had some applications that before had to spend some time performance tuning after they'd had their first cut we've moved them from that core networking into that spine leaf and they've gone in some cases in order of magnitude faster purely based on that networking topology so that was a big change for us it also allows us to much more easily scale up horizontally we can just add more spine layers and then build those up with the leaves going forward Who do you call when something goes wrong? Ha ha Do you know what all of the vendors we have? We've got a list that I'm calling the reference stack and we've done this with open stack in mind because that's the core of the project we've done each of those vendors were chosen not only for their technical capabilities and the feature sets but the culture of the teams and their ability to want to be pushed because a lot of the things we were asking them to do maybe we're a little beyond what is more traditional IT infrastructure organization would have asked so who do I call? I tend to reach out to all of them depending on the problem so it might be direct to Arista for an issue it might be directly through Red Hat our partners on the open stack side but also all of them have a very good understanding between them that if it's one of these partners' problems it's all of our problems and we've found a couple of times where we've had multiple partners swarm together to try and help us understand what the issue is and fix it and so far it's proved very very useful And is that how you guys generally talk to your customers we call any of us and we'll figure it out or do you try to have a single point of contact? Well I think the opportunity we have now with this partnership is for customers like Richard we can now deliver the option of having our HPE consulting and support services be the central point of contact One of the big value propositions that we're going to bring to the table here is rolling out an entire slew of services that include Arista that include our compute and storage and virtualization solutions so that we can actually reduce the number of those that list of numbers that a customer has to retain we can reduce that and be almost a single point of contact So as we move forward we want to really sort of drive that as one of the leading benefits of this partnership It's a big deal when you're competing against the incumbents who have large organizations for us Arista is still a start-up we're still a growing company even though we've hit a billion in revenue We don't have the consulting resources that HPE does so HPE and so I've spent a lot of my time this week here at Discover with the consulting services team I've got a few more meetings today and tomorrow and it's all about being able to deliver that total solution through HPE with the best technology and networking along with compute and storage It's a great story And your guys are comped sort of the same way they would be any channel, is that right? Yep, absolutely So we figured that out early on I think Dominic was a key to making sure that we made it completely neutral for them and our sales team is embracing the HPE team like crazy I mean, I'm on calls almost every day where we're organizing, getting the teams together and talking about which accounts do we go attack next Yeah, I mean, it's exciting, right? I mean, Cisco is such an entrenched competitor you know, you've got the Cisco trained engineers and so forth and Arista's proven, you know, consistently that you can compete, you know, in your areas a billion versus, you know, a much larger company and Dom, you guys are evolving your networking strategy considerably I mean, obviously the three comms is sort of a nice alternative but now with Aruba and the Arista partnership do you feel like that portfolio is kind of where you want it to really sort of attack the leader? Yeah, I mean, we're really set up for the next sort of decade of evolutions and transformations that are going to happen in networking I mean, you know, the Aruba acquisition was huge for us I mean, you know, that brought us the mobility capability, the mobility solutions that are really driving, you know everything that's happening in technology today because everything is mobile and so that became, you know, a very, very strong foundation and, you know, and we'd been working on this partnership for the rest of, you know, for some time because, you know, the evolution of the data center side of everything becoming cloud native and cloud enabled and everybody driving towards a secure hyper cloud model you know, we needed a sort of partnership with the best of breed solution there while still continuing to serve and service our existing customer base and those enterprises who are perhaps not ready yet and I want to stay on the sort of more traditional path so, you know, we have the sort of multiple aspects of the portfolio and as I said earlier we're going to continue to invest there but, you know, the market growth is towards that sort of, you know, cloud and software defines so it's really important for us to have a strong partnership here to drive that transformation forward for customers. All right, Richie, we're out of time but I'll close with the customer perspective Vibe on the show at Discover, what's it like for you? The vibe is good, it's also absolutely enormous it's a huge building this and it's absolutely packed so it's looking like a great show so far. It's good, a little clapping on behind us so thanks very much gentlemen for coming to theCube really appreciate your time. All right, keep it right there everybody we'll be back with our next guest this is theCube we're live from HPE Discover in London, Excel, right back.